37th EPS Conference on Plasma Physics
P1.1100
Momentum losses by charge exchange with neutral particles in H-mode
discharges at JET
T.W. Versloot1* , P.C. de Vries1,2, C. Giroud2, M. Brix2, M.G. von Hellermann1, D. Moulton3,
M.O’ Mullane4, A. Salmi5, T. Tala6, I. Voitsekhovich2, K.-D. Zastrow2 and JET-EFDA
Contributors†
JET-EFDA Culham Science Centre, Abingdon, OX14 3DB, UK
Institute Rijnhuizen, Association EURATOM-FOM, Nieuwegein, the Netherlands
2 EURATOM/CCFE Fusion Association, Culham Science Centre, Abingdon, UK
3 Imperial College of Science, Technology and Medicine, London, UK
4 Department of Physics and Applied Physics, University of Strathclyde, Glasgow, G4 0NG, UK
5 Association EURAROM-Tekes, HUT, P.O. Box 4100, 02015 TKK, Finland
6 Association EURATOM-Tekes VTT, P.P. Box 1000, 02044 VTT, Finland
1 FOM
1. Introduction
Extensive investigations both in theory and experiments have been done in recent years to
identify the influence of rotation on plasma performance. Results have shown that a radial
velocity gradient can play a role in the suppression of turbulent transport and profile stiffness [1].
It remains however largely unknown what processes determine the shape and magnitude of the
observed rotation profile. With the presence of an inwards momentum pinch [2], the edge
momentum density is observed to contribute significantly to the global confinement [3].
Therefore, in order to accurately predict the observed rotation profile, a better understanding of
the processes that determine the edge rotation is needed.
Several components play a role in momentum transport in the edge. The dominant source
of torque at JET is provided by NBI, while radial transport by outwards diffusivity and an
inwards convective pinch redistribute the momentum density. At the edge, a continuous sink is
present in the form of charge-exchange (CX) interactions between plasma ions and a neutral
particle background. Besides these direct losses, the penetration of low energy neutrals into the
plasma periphery is also believed to play a role in several plasma models related to the pedestal
shape [4] and the L-H transition [5]. In this paper, the results from a qualitative neutral transport
model are used to assess the penetration of neutral atoms into the plasma edge. A forward model
of the passive charge-exchange emission [6] is used in order to quantify the neutral density and
to estimate the magnitude of momentum and energy losses by CX interactions.
2. Experiments
A series of discharges was selected with a similar plasma configuration and equal external
heating powers, but increasing input gas flux during the flat top H-mode period. The heating and
torque deposition was majorly supplied by NBI at PTOT=15(±1)MW and TNBI=16(±1)Nm in
discharges with a plasma current of Ip=2.5MA and toroidal field of BT=2.7T. The gas-influx (Γg)
was supplied by ring valves located at the inner divertor and increased from non-fuelled to a
*
†
E-mail: [email protected]
See the Appendix of F. Romanelli et al., Proceedings of the 22nd IAEA Fusion Energy Conference, Geneva, Switzerland, 2008
37th EPS Conference on Plasma Physics
15
P1.1100
Gas scan (Ip = 2.5MA, BT = 2.5T)
0.5
Ip = 2.5MA, BT = 2.7T, Ptot = 15MW, Tφ = 16Nm
10
0.3
0.2
5
Greenwald density limit
0
0
5
10
15
20
Neutral gas influx (1021m-3s-1)
JG10.120-1c
0.1
25
0
3.2
Γ0/ne
1.27
1.24
1.26
1.35
1.76
2.55
3.3
3.4
3.5
3.6
Major radius (m)
3.7
JG10.120-2c
Thermal mach
Electron density (1019 m-3)
0.4
3.8
Figure 1: (a) Volume averaged electron density and limit versus the neutral flux through the LCFS as obtained
from the 1D neutral model. (b) Thermal mach profile (Mth=ωφR/vth) for all discharges, showing a global drop
in the mach profile with increasing neutral influx
maximum of Γg,max~4x1022 elec/s at high triangularity (δ~0.40-0.44). Immediately it was
observed that the ELM behaviour changes dramatically with external fuelling. A regular ELM
period (fELM~10Hz) is observed in the non-fuelled case, while with increased fuelling the ELM
behaviour becomes less pronounce with an increase in the baseline Dα-radiation and ELM
frequency. Higher edge electron densities (ne) are observed with higher rates of gas dosing (see
figure 1a) as also studied in [7]. The total thermal angular momentum is obtained from volume
integrating the momentum density (lφ =nimvφ) obtained using the core CXRS diagnostic. The ion
species measured is that of the main impurity, C5+, and it is assumed that all the ion species are in
equilibrium. For this analysis, the magnitude and shape of ne and Te at the pedestal is of high
importance for correctly modelling the neutral density profile as well as to forward model the
passive CX emission. High resolution Li-beam data is used in combination with Thomson
Scattering to obtain the ELM averaged profile. The electron temperature (Te) profile is obtained
from Electron Cyclotron Emission. Both ion temperature and rotation velocity are observed to
decrease as the electron density increases during the gas scan. A significant drop in thermal Mach
number over the full profile is visible in figure 1b. The angular momentum (Lφ) and total thermal
energy (Wth) decrease with gas flux under equal heating conditions. This reduction is mostly
originating from a decrease at the pedestal where also the confinement properties are degrading.
In comparison with the non-fuelled case, the fraction of angular momentum at Γg,max is 0.50
(±0.07) compared to 0.67(±0.08) for the total thermal energy (see also figure 3a). Both the
energy confinement time and momentum confinement time drop but it is interesting to note that
their ratio (Rτ= τE/τϕ) increases slightly, see figure 3b, suggesting a different behaviour between
momentum and energy losses in the presence of a neutral background.
3. Neutral Particle Dynamics
In general, the plasma is considered to be impermeable to neutrals with the penetration depth
limited to a few cm [8]. Neutral particles are unbound by the magnetic field and multiple
interactions can redistribute energy, particles and momentum. At each CX interaction, the plasma
37th EPS Conference on Plasma Physics
1015
10
ne (m-3)
n0 (m-3)
10
14
13
1012
1011
0
0.6
n0 (EDGE2D, 74312)
n0 (FRANTIC)
ne (diagn)
0.7
0.8
0.9
Norm. radius (r/a)
1.0
1.1
JG10.120-3c
1010
1019
0.4
NBI (no gas)
NBI (40% gas)
CX friction (no gas)
CX friction (40% gas)
0.3
0.2
0.1
0
0.4
0.5
0.6
0.7
0.8
0.9
1.0
0.5
0.6
0.7
0.8
Normalised radius (r/a)
0.9
1.0
0.4
0.3
0.2
0.1
0
0.4
JG10.120-4c
1020
Torque density (Nm/m3)
Pulse No: 73242 (40% gas)
Pulse No: 73244 (no gas)
Pulse No: 74312 (40% gas)
Power density (MW/m3)
1016
P1.1100
Figure 2: (a) Neutral (full) and electron (dashed) density profiles for non-fuelled and maximum fuelled
discharges. For comparison, the neutral density from a 2D simulation code (EDGE2D-EIRENE) is plotted,
showing a good agreement with the 1D neutral transport response after adjustment with PCX. (b) Beam torque
(top) and power (bottom) deposition profile and charge-exchange friction components (shown as positive)
ion and neutral swap identities. This process is considered to be a soft interaction [9] where the
outgoing neutral obtains a velocity characterised by the local ion temperature. Energy, a scalar
quantity, can be subsequently regained by the plasma in secondary interactions while momentum
transferred to the neutral fluid is assumed completely lost to either the wall or by viscous
dissipation [10]. Hence multiple CX interactions will constitute a difference in momentum and
energy losses. To determine the magnitude of CX losses, the neutral density is modelled in this
paper using a simple 1D model [11-12] by determining the neutral transport response of the
plasma. The neutral population was split in two contributions with their relative intensity and
energy obtained from visible Bremsstrahlung measurements (n0hot[0.35,10eV], n0cold[0.6,1eV]).
The modelled neutral penetration was then calibrated by forward modelling the passive chargeexchange emission (PCX) and comparing the obtained emission profile to the measured photon
flux for all mid-plane CXRS lines of sight. Figure 2a shows the adjusted n0 profiles for two
discharges as well as the poloidal averaged neutral density from a 2D edge model (EDGE2DEIRENE) for comparison [13]. With increasing neutral influx, the edge electron density increases,
shielding the core plasma from penetration from cold neutrals. The obtained neutral density is
found in both cases of similar order of magnitude at the plasma edge, n0(r/a=1.0) ~ O(1015-16)m-3.
4. Momentum and Energy Losses
Using the obtained neutral density profile, figure 2b shows the calculated charge-exchange
friction (mnin0<vσ>(vφ-v0)) and power loss (nin0<vσ>(Ti-T0)) [8]. Due to the increase in ne the
beam deposition has shifted outwards. The magnitude of CX friction drops rapidly, but is seen to
be present up to r/a~0.9 with the torque density being larger than the power loss density. The CX
losses are similar in magnitude to the local beam deposition at the edge. In figure 3a, Wth and Lφ
of all discharges are shown normalised to the non-fuelled case. A decrease is observed as the
confinement degrades but with a larger reduction in Lφ. The magnitude of CX losses increases to
approximately 10% of the total input torque with the power losses being smaller (<3%).
37th EPS Conference on Plasma Physics
JG10.120-5c
LCX, r/a = 0.95
WCX, r/a = 0.95
Lφ
Wth
1.0
0.4
0.3
0.2
0.6
0.4
0.1
1.4
0.2
0
0
Ip = 2.5MA, BT = 2.7T, Ptot = 15MW, Tφ = 16Nm
1.2
1.0
0.8
0.6
0.5
1.0
1.5
2.0
Norm. Neutral Flux (Γ0 /ne)
2.5
0
3.0
(s-1)
0.4
0
1
2
Norm. neutral flux (Γ0 /ne) (s-1)
JG10.120-6c
Norm. Lφ / Wth
0.8
1.6
Rτ
Ip = 2.5MA, BT = 2.7T, Ptot = 15MW, Tφ = 16Nm
Fraction LCX / Fraction WCX
1.2
P1.1100
3
Figure 3: (a) Wth (circles) and Lφ (triangles) normalised to a reference discharge without external fuelling.
Open symbols refer to the charge-exchange losses using the calibrated neutral density profile and integrated
up to 95% of the plasma radius (b) Confinement time ratio as a function of total normalised neutral influx
5. Conclusions
It was found that the magnitude of charge-exchange losses for momentum and energy differ
when comparing discharges at low and high gas dosing. An increase in friction losses is observed
up to 10% of the total input torque with the power losses appearing smaller. This behaviour is
consistent with the presence of multiple CX interactions within the main plasma causing
additional losses of momentum compared to energy. The uncertainty on the absolute magnitude
remains large and more detailed neutral transport calculations are required to fully quantify the
losses [14]. Also, other loss and transport processes which effect the pedestal and confinement
have been excluded, however the increase in frictional losses does suggest a direct link between
external conditions and global observed confinement. Atomic physics processes can thus play a
direct role in setting the necessary boundary conditions and therefore need to be taken into
account when predicting global plasma properties.
This work, supported by the European Communities under the contract of Association between EURATOM and
FOM and CCFE, was carried out within the framework of the European Fusion Development Agreement. The views
and opinions expressed herein do not necessarily reflect those of the European Commission.
References
[1] Mantica P. et al.,Phys. Rev. Lett., 102, 175002, (2009)
[2] Tala T., et al., Plasma Phys. Control. Fusion, 49, B291
(2007)
[3] de Vries P.C. et al., Plasma Phys. Control Fusion, 52,
065004 (2010)
[4] Mahdavi, et. el, Phys. Plasmas, 10, p3984 (2003)
[5] Owen L.W., et. al, Plasma. Phys. Control. Fusion, 40,
p717 (1998)
[6] Tunklev M., et. al., Plasma Phys. Control. Fusion, 41,
985-1004 (1999)
[7] Maddison G., et al. Proc. 19th PSI Conference, San
Diego, 2010
[8] Carreras B.A., et al. Phys. Plasmas, 3, 11, 4106-4114
(1996)
[9] Tendler M., et al., Fusion Technology, 11, 289-310
(1987)
[10] Hazeltine R.D., et al., Nucl. Fusion, 32, 3-14 (1992)
[11] S. Tomar, Journal of Comp. Phys. 40, 104-119,
(1981)
[12] Valovic M., et al., Plasma Phys. Control. Fusion,
46, 1877-1889 (2004)
[13] Moulton D., et al., Proc. 19th PSI Conference, San
Diego, 2010
[14] Friis Z.W., et al., Phys. Plasmas, 17, 022507 (2010)